CN106599846B - A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition - Google Patents
A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition Download PDFInfo
- Publication number
- CN106599846B CN106599846B CN201611157679.0A CN201611157679A CN106599846B CN 106599846 B CN106599846 B CN 106599846B CN 201611157679 A CN201611157679 A CN 201611157679A CN 106599846 B CN106599846 B CN 106599846B
- Authority
- CN
- China
- Prior art keywords
- color
- annulus
- traffic
- computer vision
- easy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/582—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of traffic signs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Abstract
A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition, includes the following steps, is added on existing label ontology by multiple monochromatic annulus concentric polychrome annulus patterns that relationship is arranged in a combination by size;It include the image sequence of concentric polychrome annulus by video camera acquisition;Linear scanning on transverse and longitudinal direction is carried out to collected every frame image, and forms a team to form color vector array according to the dithering pixel in predefined color set;The all colours vector array filtered out is analyzed, a color vector array is ultimately generated;According to final color vector array, road signs information corresponding to matched defined traffic sign coding mode is found.Simple, the accurate identification to traffic sign may be implemented in the present invention, influence of the disturbing factor to detection and identification accuracy can be reduced, the safety and current high efficiency that can help to guarantee road traffic, especially have a good application prospect to exploitation DAS (Driver Assistant System).
Description
Technical field
The present invention relates to intelligent transportation, auxiliary driving field, and in particular to a kind of traffic for being easy to Computer Vision Recognition
The recognition methods of sign board.
Background technique
One of the key problem that automotive safety auxiliary drives is exactly real-time perception road signs information.In vehicle driving
In the process, acquisition, detection and identification promptly and accurately are carried out to a variety of traffic signs occurred in road environment, can be driven for auxiliary
The road environment information of system offer promptly and accurately is provided, thereby may be ensured that the safety and current high efficiency of road traffic.
And Computer Automatic Recognition system is one of the important technology that auxiliary drives.
Traffic mark board is the important carrier for carrying the information such as traffic guide, alarm and behavior limitation, meanwhile, it has become
The important component of intelligent transportation.As a kind of public identifier, traffic sign has significant color and shape feature, and
It has been a very important research direction using vehicle-mounted automatic recognition system identification road signs.It is automatic in traffic sign
In identifying system, the Traffic Sign Images under traffic environment are acquired generally by the video camera being installed on vehicle, are then made
Traffic sign is identified and understood with the detection of computer vision and identification technology.However, the traffic under natural scene
There are a variety of disturbing factors for sign image, this has seriously affected the accurate detection and identification of traffic sign, these disturbing factors master
Caused by the position deviation of interference and installation caused by interference, weather conditions caused by the interference to be had powerful connections, illumination variation
Interference etc..
Summary of the invention
In view of the above existing problems in the prior art, the present invention provides a kind of traffic signs for being easy to Computer Vision Recognition
The recognition methods of board, this method accurately can detect and identify the information on traffic mark board, can reduce disturbing factor to detection
With the influence of identification accuracy, the safety for guaranteeing road traffic and current high efficiency can help to.
To achieve the goals above, the present invention provides a kind of identification side of traffic mark board for being easy to Computer Vision Recognition
Method includes the following steps,
Step 1: auxiliary sign is added at the edge of label ontology identical with existing traffic mark board and gap
To form the traffic sign for being easy to Computer Vision Recognition, the auxiliary sign is by convenient for by video camera multiple monochromes collected
Relationship arranges annulus by size, and this multiple monochromatic annulus generates a concentric polychrome annulus, the multiple monochrome annulus
Color and multiple monochromatic putting in order for annulus represent specific road signs information;Monochromatic annulus in the auxiliary sign
Quantity C be definite value, and the color of adjacent monochromatic annulus is different, wherein the color used selected by each monochrome annulus is derived from
Predefined color set S, and in the set S any two color color space vector Euclidean distance be greater than threshold value T1;
Step 2: pass through the image sequence of auxiliary sign of the video camera acquisition comprising concentric polychrome annulus;
Step 3: every frame image that size is N × M is pre-processed, removes or weakens noise jamming, increase image
By force;
Step 4: carrying out the linear scanning on horizontal and vertical to each frame image, for each straight line scanned,
Correspondingly initialize a color vector array Ai;
Step 5: along the pixel p on every straight line l successively scanned straight linesiIf color value s ∈ S is scanned, and | | pi-s
||2≤ T2, then the pixel p is recordediCorresponding predefined color value, wherein T2For threshold value, and T2< T1;
Step 6: next pixel p on straight line l is continued to scan onjIf certain pixel p is calculated by step 5jIt is corresponding
Predefined color valueS is then stored in color vector array Ai, step 5 is then repeated, step 6 is otherwise executed;
Step 7: after scanning through the pixel on straight line l, if by the color vector array A of the linear structureiMiddle element
Quantity is more than or less than the quantity C of determining monochromatic annulus, then abandons color vector array Ai;
Step 8: after scanning through a frame image, all colours vector array A retained is analyzedi, ultimately generate a face
Color vector array A, the element value of array A are derived from the color value that array element is taken under same index in most of arrays;
Step 9: according to final color vector array A, matched defined traffic sign coding mould is found
Formula, each encoding model correspond to a kind of road signs information, to complete to identify.
Median filtering method or low-pass filtering can be used as a preference, removing in the step 3 or weakening noise jamming
Method.
Further, in order to accelerate scanning speed, horizontal and vertical straight line is carried out to each frame image in the step 4 and is swept
When retouching, using being scanned by the way of a row or column.
As a preference, in the step 4 to each frame image carry out horizontal and vertical linear scanning when, using every
Multirow is scanned every the mode of multiple row.
As a preference, traffic sign coding mode is constructed by following methods in the step 9, firstly, choosing certain
The color of quantity is as predefined color collection, then, concentrates a certain number of colors of selection to carry out arrangement group from predefined color
Cooperation is the coding of some traffic sign.
Further, in order to accelerate matching speed, the traffic sign coding found in the step 9 is by using tree-shaped knot
It is chosen in the traffic sign coded set of structure tissue.
The present invention has installed a kind of concentric polychrome circle being made of multiple monochromatic annulus additional on the basis of existing traffic sign
Ring is easy to by the auxiliary sign of Computer Vision Recognition, due in actual traffic environment, existing traffic sign be easy because
It the situations such as blocks, tilt, rotate and obscures, computer recognition system is caused to be lower the identification of traffic sign.And it is of the invention
In auxiliary sign be multiple monochromatic concentric loops compositions, acquired by video camera and after pretreatment, by lateral and vertical
To the linear each frame image of scanning, and whether the pixel analyzed on each frame image belongs to predefined color set, often sweeps
Retouch straight line and initialize a color vector array, finally formed multiple vector arrays by analysis obtain one it is final
Vector array matches the road signs information that this kind of auxiliary sign is carried by the final vector array.Thus, the party
Method accurately can detect and identify the information on traffic mark board, can reduce disturbing factor to the shadow of detection and identification accuracy
It rings, can help to the safety for guaranteeing road traffic and current high efficiency.In addition, using multiple monochromatic annulus according to certain
The concentric polychrome annulus pattern of arrangement mode group indicates that certain information can also be managed under different literals background by driver
Solve and recognize reading.Meanwhile the auxiliary sign is added to the edge of existing traffic mark board and gap nor affects on driver
Identification to existing normal traffic mark.The recognition methods base of traffic mark board with auxiliary sign proposed by the invention
In computer vision technique, for developing automotive safety DAS (Driver Assistant System), guarantee driving safety is of great significance.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of traffic mark board in the present invention;
Fig. 2 is the structural schematic diagram of auxiliary sign in the present invention;
Fig. 3 is the flow chart of recognition methods in the present invention.
Specific embodiment
The present invention will be further explained below with reference to the attached drawings.
As shown in figures 1 and 3, a kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition, including it is following
Step,
Step 1: auxiliary sign is added at the edge of label ontology identical with existing traffic mark board and gap
To form the traffic sign for being easy to Computer Vision Recognition, as shown in Fig. 2, the auxiliary sign is acquired by being convenient for by video camera
Multiple monochromatic annulus relationship arranges by size, and this multiple monochromatic annulus generates a concentric polychrome annulus, described more
The color of a monochrome annulus and multiple monochromatic putting in order for annulus represent specific road signs information;In the auxiliary sign
The quantity C of monochromatic annulus be definite value, and the color of adjacent monochromatic annulus is different, uses wherein each monochrome annulus is selected
Color be derived from predefined color set S, and any two color is big in the Euclidean distance of color space vector in set S
In threshold value T1;
Step 2: pass through the image sequence of auxiliary sign of the video camera acquisition comprising concentric polychrome annulus;
Step 3: every frame image that size is N × M is pre-processed, removes or weakens noise jamming, increase image
By force;
Step 4: carrying out the linear scanning on horizontal and vertical to each frame image, for each straight line scanned,
Correspondingly initialize a color vector array Ai;
Step 5: along the pixel p on every straight line l successively scanned straight linesiIf color value s ∈ S is scanned, and | | pi-s
||2≤ T2, then the pixel p is recordediCorresponding predefined color value, wherein T2For threshold value, and T2< T1;
Step 6: the lower pixel p of straight line l is continued to scan onjIf certain pixel p is calculated by step 5jCorresponding is pre-
Define color valueS is then stored in color vector array Ai, step 5 is then repeated, step 6 is otherwise executed;
Step 7: after scanning through the pixel on straight line l, if by the color vector array A of the linear structureiMiddle element
Quantity is more than or less than the quantity C of determining monochromatic annulus, then abandons color vector array Ai;
Step 8: after scanning through a frame image, all colours vector array A retained is analyzedi, ultimately generate a face
Color vector array A, the element value of array A are derived from the color value that array element is taken under same index in most of arrays;I.e.
The element value of each position is chosen by mode below in array A, all color vector array A that will be retainediPhase
Element value with position is placed in a comparative group and is compared, by that most element of quantity identical in the comparative group
Value takes out the element value as the position in array A.
Step 9: according to final color vector array A, matched defined traffic sign coding mould is found
Formula, each encoding model correspond to a kind of road signs information, to complete to identify.
Median filtering method or low-pass filtering can be used as a preference, removing in the step 3 or weakening noise jamming
Method.
It is adopted to accelerate scanning speed when carrying out horizontal and vertical linear scanning to each frame image in the step 4
It is scanned with the mode every a row or column.
It is adopted to accelerate scanning speed when carrying out horizontal and vertical linear scanning to each frame image in the step 4
It is scanned with every multirow or every the mode of multiple row.
As a preference, traffic sign coding mode is constructed by following methods in the step 9, firstly, choosing certain
The color of quantity is as predefined color collection, then, concentrates a certain number of colors of selection to carry out arrangement group from predefined color
Cooperation is the coding of some traffic sign.
In order to accelerate matching speed, the traffic sign coding found in the step 9 is by using tree tissue
It is chosen in traffic sign coded set.
The present invention has installed a kind of concentric polychrome circle being made of multiple monochromatic annulus additional on the basis of existing traffic sign
Ring is easy to by the auxiliary sign of Computer Vision Recognition, due in actual traffic environment, existing traffic sign be easy because
It the situations such as blocks, tilt, rotate and obscures, computer recognition system is caused to be lower the identification of traffic sign.And it is of the invention
In auxiliary sign be multiple monochromatic concentric loops compositions, acquired by video camera and after pretreatment, by lateral and vertical
To the linear each frame image of scanning, and whether the pixel analyzed on each frame image belongs to predefined color set, often sweeps
Retouch straight line and initialize a color vector array, finally formed multiple vector arrays by analysis obtain one it is final
Vector array matches the road signs information that this kind of auxiliary sign is carried by the final vector array.Thus, the party
Method accurately can detect and identify the information on traffic mark board, can reduce disturbing factor to the shadow of detection and identification accuracy
It rings, can help to the safety for guaranteeing road traffic and current high efficiency.In addition, using multiple monochromatic annulus according to certain
The concentric polychrome annulus pattern of arrangement mode group indicates that certain information can also be managed under different literals background by driver
Solve and recognize reading.The recognition methods of traffic mark board with auxiliary sign proposed by the invention is based on computer vision technique,
It is of great significance for developing automotive safety DAS (Driver Assistant System), guarantee driving safety.
Claims (6)
1. a kind of recognition methods for the traffic mark board for being easy to Computer Vision Recognition, which is characterized in that include the following steps,
Step 1: auxiliary sign is added with shape at the edge of label ontology identical with existing traffic mark board and gap
At the traffic sign for being easy to Computer Vision Recognition, the auxiliary sign is by convenient for by video camera multiple monochromatic annulus collected
Relationship arranges by size, and this multiple monochromatic annulus generates a concentric polychrome annulus, the face of the multiple monochrome annulus
Color and multiple monochromatic putting in order for annulus represent specific road signs information;The number of monochromatic annulus in the auxiliary sign
Amount C is definite value, and the color of adjacent monochromatic annulus is different, wherein the color used selected by each monochrome annulus be derived from it is predetermined
Justice color set S, and in the set S any two color color space vector Euclidean distance be greater than threshold value T1;
Step 2: pass through the image sequence of auxiliary sign of the video camera acquisition comprising concentric polychrome annulus;
Step 3: every frame image that size is N × M is pre-processed, removes or weakens noise jamming, make image enhancement;
Step 4: the linear scanning on horizontal and vertical is carried out to each frame image, for each straight line scanned, accordingly
Ground initializes a color vector array Ai;
Step 5: along the pixel p on every straight line l successively scanned straight linesiIf color value s ∈ S is scanned, and | | pi-s||2<
=T2, then the pixel p is recordediCorresponding predefined color value, wherein T2For threshold value, and T2< T1;
Step 6: next pixel p on straight line l is continued to scan onjIf certain pixel p is calculated by step 5jCorresponding is pre-
Define color valueS is then stored in color vector array Ai, step 5 is then repeated, step 6 is otherwise executed;
Step 7: after scanning through the pixel on straight line l, if by the color vector array A of the linear structureiThe quantity of middle element
More than or less than the quantity C of determining monochromatic annulus, then color vector array A is abandonedi;
Step 8: after scanning through a frame image, all colours vector array A retained is analyzedi, ultimately generate a color to
Array A is measured, the element value of array A is derived from the color value that array element is taken under same index in most of arrays;
Step 9: according to final color vector array A, matched defined traffic sign coding mode is found, often
A kind of encoding model corresponds to a kind of road signs information, to complete to identify.
2. a kind of recognition methods of traffic mark board for being easy to Computer Vision Recognition according to claim 1, feature
It is, removing in the step 3 or weakening noise jamming can be used median filtering method or low pass filtering method.
3. a kind of recognition methods of traffic mark board for being easy to Computer Vision Recognition according to claim 1, feature
It is, when carrying out horizontal and vertical linear scanning to each frame image in the step 4, using by the way of a row or column
It is scanned.
4. a kind of recognition methods of traffic mark board for being easy to Computer Vision Recognition according to claim 1, feature
In when carrying out horizontal and vertical linear scanning to each frame image in the step 4, using every multirow or by the way of multiple row
It is scanned.
5. a kind of identification side of traffic mark board for being easy to Computer Vision Recognition according to any one of claims 1 to 4
Method, which is characterized in that traffic sign coding mode is constructed by following methods in the step 9, firstly, choosing a certain number of
Color is as predefined color collection, then, concentrates from predefined color and chooses a certain number of colors progress permutation and combination conducts
The coding of some traffic sign.
6. a kind of recognition methods of traffic mark board for being easy to Computer Vision Recognition according to claim 5, feature
It is, the traffic sign coding found in the step 9 selects in the traffic sign coded set by using tree tissue
It takes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611157679.0A CN106599846B (en) | 2016-12-15 | 2016-12-15 | A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611157679.0A CN106599846B (en) | 2016-12-15 | 2016-12-15 | A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106599846A CN106599846A (en) | 2017-04-26 |
CN106599846B true CN106599846B (en) | 2019-09-13 |
Family
ID=58801989
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611157679.0A Active CN106599846B (en) | 2016-12-15 | 2016-12-15 | A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106599846B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107705347A (en) * | 2017-09-12 | 2018-02-16 | 方家齐 | One Trademark design accessory system |
CN107844761B (en) * | 2017-10-25 | 2021-08-10 | 海信集团有限公司 | Traffic sign detection method and device |
CN111539424A (en) * | 2020-04-21 | 2020-08-14 | 北京云从科技有限公司 | Image processing method, system, device and medium based on OCR |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1102212A2 (en) * | 1999-11-19 | 2001-05-23 | Lg Electronics Inc. | Colour image quantization system |
CN101201903A (en) * | 2006-12-12 | 2008-06-18 | 佳能株式会社 | Image processing apparatus, method for controlling image processing apparatus |
CN102364560A (en) * | 2011-10-19 | 2012-02-29 | 华南理工大学 | Traffic sign convenient for electronic identification and method for identifying traffic sign |
CN102800243A (en) * | 2012-07-18 | 2012-11-28 | 湖南大学科技园有限公司 | Anti-counterfeiting annular code and encoding method thereof |
CN102831420A (en) * | 2012-08-17 | 2012-12-19 | 银江股份有限公司 | Circular traffic sign positioning method based on color information and randomized circle detection |
CN103065494A (en) * | 2012-04-12 | 2013-04-24 | 华南理工大学 | Free parking space detection method based on computer vision |
CN103544484A (en) * | 2013-10-30 | 2014-01-29 | 广东工业大学 | Traffic sign identification method and system based on SURF |
CN104504904A (en) * | 2015-01-08 | 2015-04-08 | 杭州智诚惠通科技有限公司 | Mobile acquisition method of transportation facility |
CN104636779A (en) * | 2013-11-11 | 2015-05-20 | 覃政 | Annular code identifier recognition system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6266442B1 (en) * | 1998-10-23 | 2001-07-24 | Facet Technology Corp. | Method and apparatus for identifying objects depicted in a videostream |
-
2016
- 2016-12-15 CN CN201611157679.0A patent/CN106599846B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1102212A2 (en) * | 1999-11-19 | 2001-05-23 | Lg Electronics Inc. | Colour image quantization system |
CN101201903A (en) * | 2006-12-12 | 2008-06-18 | 佳能株式会社 | Image processing apparatus, method for controlling image processing apparatus |
CN102364560A (en) * | 2011-10-19 | 2012-02-29 | 华南理工大学 | Traffic sign convenient for electronic identification and method for identifying traffic sign |
CN103065494A (en) * | 2012-04-12 | 2013-04-24 | 华南理工大学 | Free parking space detection method based on computer vision |
CN102800243A (en) * | 2012-07-18 | 2012-11-28 | 湖南大学科技园有限公司 | Anti-counterfeiting annular code and encoding method thereof |
CN102831420A (en) * | 2012-08-17 | 2012-12-19 | 银江股份有限公司 | Circular traffic sign positioning method based on color information and randomized circle detection |
CN103544484A (en) * | 2013-10-30 | 2014-01-29 | 广东工业大学 | Traffic sign identification method and system based on SURF |
CN104636779A (en) * | 2013-11-11 | 2015-05-20 | 覃政 | Annular code identifier recognition system |
CN104504904A (en) * | 2015-01-08 | 2015-04-08 | 杭州智诚惠通科技有限公司 | Mobile acquisition method of transportation facility |
Also Published As
Publication number | Publication date |
---|---|
CN106599846A (en) | 2017-04-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102708356B (en) | Automatic license plate positioning and recognition method based on complex background | |
CN102375982B (en) | Multi-character characteristic fused license plate positioning method | |
CN111104903B (en) | Depth perception traffic scene multi-target detection method and system | |
CN102722707A (en) | License plate character segmentation method based on connected region and gap model | |
CN105488492B (en) | A kind of color image preprocess method, roads recognition method and relevant apparatus | |
CN101154271A (en) | License plate character segmentation method based on fast area labeling algorithm and license plate large-spacing locating method | |
CN101739566B (en) | Self-adapting projection template method-based automobile plate positioning method | |
CN106599846B (en) | A kind of recognition methods for the traffic mark board being easy to Computer Vision Recognition | |
CN105608429B (en) | Robust method for detecting lane lines based on difference excitation | |
CN103870803A (en) | Vehicle license plate recognition method and system based on coarse positioning and fine positioning fusion | |
CN103065138A (en) | Recognition method of license plate number of motor vehicle | |
CN106570494A (en) | Traffic signal lamp recognition method and device based on convolution neural network | |
CN109448000B (en) | Segmentation method of traffic direction sign image | |
CN103903018A (en) | Method and system for positioning license plate in complex scene | |
CN104715252A (en) | License plate character segmentation method with combination of dynamic template and pixel points | |
Yang et al. | A vehicle license plate recognition system based on fixed color collocation | |
CN100385452C (en) | Registration number character dividing method | |
CN103488978A (en) | License plate location method based on gray level jump and character projection interval mode | |
CN103984963B (en) | Method for classifying high-resolution remote sensing image scenes | |
CN104200207A (en) | License plate recognition method based on Hidden Markov models | |
CN106529472B (en) | Object detection method and device based on large scale high-resolution high spectrum image | |
CN111160205A (en) | Embedded multi-class target end-to-end unified detection method for traffic scene | |
CN103324958A (en) | License plate locating method based on projection method and SVM under complex background | |
CN106919939A (en) | A kind of traffic signboard Tracking Recognition method and system | |
KR20100064140A (en) | Method for detecting traffic sign board in intelligent vehicle and system for executing the method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |